Generative AI as a Credibility Scaffold in Peer Assessment: Evidence from a Chinese Undergraduate Course

• In Event: AERA Poster Session Thursday 9:45am
  In Poster Session: Division H Poster Session: Assessment in Schools

Abstract

This study employs a mixed-methods design to explore how generative artificial intelligence (GAI) can enhance the credibility of peer assessment in higher education. An 18-week blended course with 27 undergraduates implemented a GAI-supported model that generated credibility scores based on feedback clarity, constructiveness, and relevance. Results showed strong alignment between GAI and instructor ratings, validating GAI’ s credibility evaluation capacity. GAI scores also revealed significant differences between high- and low-performing students, demonstrating its ability to detect variation in cognitive and evaluative performance. Questionnaire data indicated strong student acceptance, particularly in self-efficacy and perceived usefulness. These findings provide preliminary evidence that GAI-supported models can enable fair, scalable assessment while fostering deeper peer interaction, reflection, and engagement in technology-enhanced learning environments.

Authors

• Lin Dai, East China Normal University

• Jing Leng, East China Normal University

• Honghuan Lu, East China Normal University